Broughton, R.K. et al

UKCEH Land Cover Plus: Hedgerows 2016-2021 (England)

https://doi.org/10.5285/d90a3733-2949-4dfa-8ac2-a88aef8699be Cite this dataset

This dataset contains a model of the extent and height class of woody linear features on field boundaries in England, including hedgerows, tree lines and semi-natural thickets of shrubs and trees. The model was derived from processing of the Environment Agency lidar product (National Lidar Programme), captured in 2016-2021, and the linework is consistent with the polygon boundaries used in the UK Centre for Ecology & Hydrology (UKCEH) Land Cover Map (LCM).

Due to expected low densities of woody linear features, and potential uncertainty, areas excluded from the coverage include LCM classes of mountain/moor/heath, open water and the coastal zone, urban/suburban areas and woodlands (i.e., areas of continuous non-linear woody cover). An area of approximately 24 x 25 km on the North Yorkshire Moors was also excluded from the model, as this area was not mapped by the EA lidar.

Publication date: 2024-01-16

Format

GeoPackage

Spatial information

Study area
Spatial representation type: Vector
Spatial reference system: OSGB 1936 / British National Grid

Temporal information

Temporal extent: 2016-01-01 to 2021-12-31

Provenance & quality

This model was produced by collating Environment Agency airborne lidar data, Land Cover Map (LCM) linear data, and Ordnance Survey (OS) data. The lidar data were filtered and pixels were removed based on their attributes and coincidence with OS data (masking). Vectorized output was used to classify a linear framework of the LCM, creating a classified polyline model of linear features.

The model was validated by ‘ground-truthing’ against 38 test squares of 1 km^2 each, which were surveyed on the ground by trained fieldworkers in 2022 using Countryside Survey methods. The results showed good agreement between the lidar model and the ground-truthing, allowing for a spatial buffer tolerance around features to account for the different linear frameworks they were based upon. At the summary scale across all test squares, the total woody feature lengths showed a 96% agreement. At the scale of individual test squares, there was an average 76% agreement of woody feature lengths within a 20 m buffer tolerance. Height classifications of the model showed an exact agreement for 32% of the total feature length, and 63% agreement within a tolerance of one class above or below.